def test_lr_scheduler(self):
data_set, model = prepare_env()
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)
trainer = Trainer(data_set, model, optimizer=optimizer, loss=BCELoss(pred="predict", target="y"), batch_size=32,
n_epochs=5, print_every=50, dev_data=data_set,
metrics=AccuracyMetric(pred="predict", target="y"), use_tqdm=False,
callbacks=[LRScheduler(torch.optim.lr_scheduler.StepLR(optimizer, step_size=10, gamma=0.1))],
check_code_level=2)
trainer.train()
embed = StackEmbedding([word_embed, char_embed])
model = CNNBiLSTMCRF(embed,
hidden_size=1200,
num_layers=1,
tag_vocab=data.vocabs[Const.TARGET],
encoding_type=encoding_type,
dropout=dropout)
callbacks = [
GradientClipCallback(clip_value=5, clip_type='value'),
EvaluateCallback(data.datasets['test'])
]
optimizer = SGD(model.parameters(), lr=lr, momentum=0.9)
scheduler = LRScheduler(
LambdaLR(optimizer, lr_lambda=lambda epoch: 1 / (1 + 0.05 * epoch)))
callbacks.append(scheduler)
trainer = Trainer(train_data=data.get_dataset('train'),
model=model,
optimizer=optimizer,
sampler=BucketSampler(num_buckets=100),
device=0,
dev_data=data.get_dataset('dev'),
batch_size=batch_size,
metrics=SpanFPreRecMetric(
tag_vocab=data.vocabs[Const.TARGET],
encoding_type=encoding_type),
callbacks=callbacks,
num_workers=1,
n_epochs=100,
requires_grad=False,
normalize=False),
ElmoLayer=None,
args_of_imm={
"input_size": 300,
"hidden_size": arg.hidden_size,
"dropout": arg.dropout,
"use_allennlp": False,
},
)
optimizer = Adadelta(lr=arg.lr, params=model.parameters())
scheduler = StepLR(optimizer, step_size=10, gamma=0.5)
callbacks = [
LRScheduler(scheduler),
]
if arg.task in ['snli']:
callbacks.append(
FitlogCallback(data_info.datasets[arg.testset_name], verbose=1))
elif arg.task == 'mnli':
callbacks.append(
FitlogCallback(
{
'dev_matched': data_info.datasets['dev_matched'],
'dev_mismatched': data_info.datasets['dev_mismatched']
},
verbose=1))
trainer = Trainer(train_data=data_info.datasets['train'],
